Contact centres are widely used to allow an enterprise or business to efficiently handle customer enquiries, complaints and support, and to allow an enterprise to make contact with existing or potential customers.
A widely recognised difficulty exists when contact centres cannot efficiently handle the volume of calls at any given time. (The term “calls” here is not intended to denote solely voice telephony calls as many contact centres provide multimedia communication between agents of the centre and customers or remote users, so that “calls” is intended to encompass voice and video calls, Internet chat sessions, emails, short messaging service (SMS) messages and any other such communications between a remote user and an agent. Similarly, while the term “call centre” will be used interchangeably with “contact centre”, it is to be understood that this is not a limitation on the type of contact centre to which the invention applies.)
A typical contact centre will have two main components, namely a network interface such as a switch or a private branch exchange (PBX) which handles the calls themselves, and a management component such as a contact centre server for controlling the operation of the PBX (the PBX and the management component can in fact be the same physical implementation) to direct calls to suitable agents. Thus, based on information gleaned from the call in its initial stages (e.g. from menu choices made in an interactive voice response (IVR) session), the contact centre server will determine a skillset which an agent would ideally require to handle the call. An agent is identified from those currently active at the centre, taking into account the number of calls currently on hold for that agent and other relevant data available to the contact centre server, and the PBX is directed to transfer the call to the directory number (DN) or the position ID of that agent by the contact centre server when the agent becomes free.
Nortel Networks provides a contact centre solution in which its Meridian PBX is associated with a Symposium Call Centre Server (SCCS). (Nortel, Meridian, Symposium and SCCS are trade marks.) The possibility currently exists to network a number of contact centres by providing an Internet (or other WAN) link between the respective SCCS units of different contact centres, and by providing telephony links (either dedicated ISDN (integrated services digital network) lines or links over the public switched telephone network (PSTN)) between the respective Meridian PBX units. A Networked Call Centre (NCC) server is also provided on the WAN to propagate to each SCCS the network address and routing tables for all of the other networked call centres.
Each SCCS unit therefore registers with the NCC server, and the NCC server defines and propagates the routing tables for each node to each other node. The NCC server then drops out of the process (though it may stay connected to the SCCS units to collect statistical information).
When a SCCS determines that it cannot efficiently handle a call to its associated PBX (either due to a lack of agents with the needed skillset, or to the centre being too busy), it may broadcast a message to each of the other SCCS units to determine if any other contact centre can take the call. If and when a response is received, the SCCS at the call centre to which the call was made instructs the Meridian PBX to transfer the call to the remote contact centre over the dedicated or public telephony link.
While this solution does allow a network of contact centres to make use of an overcapacity in one centre to relieve an overload at another centre, it is by no means an ideal solution. Among the problems with this solution are:
1. The first centre to reply will normally be allocated the call. Depending on the speed of network traffic between the remote centres, this node may simply receive the call because it is the first to reply, whereas a more suitable centre may be slower to respond and thus will not be allocated the call.2. A remote centre may claim a contact to increase its turnover, but the call might be placed in a long queue after it is transferred to the remote centre, which leads to a worse overall service for the customer.3. Conversely a contact centre might try to offload more than its fair share of calls, and transfer a call to a busier centre where the customer has a longer wait.4. There is no way of telling which agent the call is to be transferred to, and the agent must be assigned to the Networked Skillset. If the originating centre is attempting to provide the best possible service to its own customers, it may therefore be reluctant to hand a call over to a remote centre.
U.S. Pat. No. 6,115,693 (McDonough et al., assigned to Andersen Consulting LLP) describes a virtual sales and service centre in which a plurality of call centres are connected to a “quality centre” which acts as a management point for all of the call centres. Calls are received centrally by the quality centre and then are distributed out to the individual call centres. In this system, all of the call centres are subordinate to the quality centre which tightly controls the operation of each centre. The intended result of this is that the network of call centres is presented as a unified resource to the customer.
U.S. Pat. No. 6,373,836 (Deruygin et al., assigned to Genesys Telecommunications Laboratories, Inc.) describes an IP telephony network centre which has a plurality of call centres managed by a central controller. The central controller accepts IP telephony calls and routes the calls to the call centres for distribution to agents of those centres. The call centres report statistical information to the central controller via a dedicated data link in such a way that in case of failure of a data link, the controller can still make educated guesses as to the status of the call centre.
In both the system of U.S. Pat. No. 6,115,693 and that of U.S. Pat. No. 6,373,836, the calls are centrally received and distributed. The individual call centres are tightly coupled to the central unit. While U.S. Pat. No. 6,373,836 contemplates the failure of a dedicated data link, neither system could function in the absence of the central unit, since this unit makes all of the decisions as to how calls should be routed.
Furthermore, because the networked centres are so tightly bound to the central unit, the systems are not designed to accommodate casual or freelance call centres being used to handle call overloads.
Cisco Systems provide a system called “Intelligent Contact Management” which is intended to manage and transfer contacts between different sites on a corporation. However, this product requires dedicated links between sites and is therefore unsuitable for use in integrating third-party or contract call centres into the network temporarily.
In busy contact centre networks, there may be, for example, 40,000 or more contacts received per hour. Whether these contacts are analysed and routed at the nodal (contact centre) level, or at the network management unit level, there is a significant processing overhead in analysing the contacts.
It is an object of the present invention to address the above disadvantages and provide a system and method of managing call centres which is more flexible in allowing calls to be transferred between independent contact centres than is the case in systems and methods known from the prior art.